Gemini Spark could be Google’s plan for an autonomous employee

A new onboarding screen is seen x user fundu The Gemini web app reveals something called “Gemini Spark” (beta), which is an always-on AI agent that handles things in the background instead of waiting for users to type one at a time (via test list).

The onboarding text says the AI ​​agent can pull data from linked apps, chat history, scheduled tasks, websites you’ve logged in to, location data, and personal intelligence. That context will enable Spark to deal with inboxes, end online workflows, and potentially automate multi-step tasks across Google services and third-party platforms.

Google has previously experimented with AI agents with internal projects like Remi, but the current Gemini Agent tools are largely limited to paying AI Ultra users. The Spark appears to be the next step in that evolution, but this time with a broader consumer-facing angle.

Meanwhile, Google is also warning users about the risks. The onboarding screen says Gemini Spark may share sensitive information with third parties or make purchases without verification in some situations. The system can also store remote browser session data, including login information and remote execution status, so that workflows can continue running in the background. Gemini Settings will apparently allow users to clear that data or turn off connected services.

Competitors such as OpenAI and Anthropic are also creating agent-style systems that can actively surf the web, perform tasks, and manage workflows with minimal user input. Tools are being designed to act on behalf of users rather than simply answer questions.

The leak also mentions something called “skills”, which hints that Spark could take advantage of modular templates for tasks or app integration to expand its capabilities over time. If these leaks are true, it could mean that Gemini Spark is one of the major announcements at Google I/O on May 19.